Solutions

Challenge I

1. Import bike-share excel file, then put the last_updated column to a readable datetime value.

1. Import excel file

library(readxl)
BShare_xls <- read_excel('bike-share.xlsx')
#str(BShare_xls) # alternative to function head

2. Readable Last_updated column

BShare_xls$last_updated<-as.POSIXct(BShare_xls$last_updated,origin = '1970-01-01')
head(BShare_xls,3)

2. Using the iot excel dataset, modify the device column by removing the colon that separate each term (for example 1c:bf:ce:15:ec:4d becomes 1cbfce15ec4d).

1. Import excel file

iot1<-read_excel('iot.xlsx')
head(iot1,3)

2. Modify device column

library(stringr)
library(purrr)
library(rex)

## 
## Attaching package: 'rex'

## The following object is masked from 'package:stringr':
## 
##     regex

iot1<-modify_at(iot1,'device',~str_split(.,':')) # applies the split function to the column device
new<-list()
for (i in 1:nrow(iot1)){ 
  # paste combines the strings into one phrase 
  val<-paste(iot1$device[[i]][1],iot1$device[[i]][2],iot1$device[[i]][3],iot1$device[[i]][4],
         iot1$device[[i]][5],iot1$device[[i]][6])
  # gsub removes space between terms 
  val<-gsub("\\s+", "", val, perl=TRUE)
  new<-append(new,val)
}
iot1$device<- unlist(new)
head(iot1,3)

Challenge II

1. Remove the columns last_updated, id and ttl from the bike share excel dataset.

BShare_xls<-BShare_xls[,4:ncol(BShare_xls)]
# or
#BShare_xls<-BShare_xls[,names(BShare_xls)!=c('last_updated','ttl','id')]
head(BShare_xls,3)

2. Merge the bike-share.xlsx and iot.xlsx depending on time

library(dplyr)

## 
## Attaching package: 'dplyr'

## The following object is masked from 'package:rex':
## 
##     matches

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

BShare_xls$ts<-BShare_xls$rec_update
BS_iot<-inner_join(BShare_xls,iot1,by='ts')
# another method
BS_iot<-merge(BShare_xls,iot1,by='ts')
#head(BS_iot,3)

3. Create a new dataset which is the subset of the one in 2. where the co value is greater than the mean value of the iot.xlsx dataset.

BS_iot_m<-BS_iot[which(BS_iot$co > mean(iot1$co)),]
# or
#BS_iot_m<-subset(BS_iot, co> mean(iot1$co))
#dim(BS_iot_m)

Challenge III:

1. Using iot_telemetry dataset, visualize the distribution of smoke of each device.

1. read the csv data

iot_tel<-read.csv('iot_telemetry_data.csv')
head(iot_tel,3)

2. Visualize the distribution

library(plotly)

## Loading required package: ggplot2

## 
## Attaching package: 'plotly'

## The following object is masked from 'package:ggplot2':
## 
##     last_plot

## The following object is masked from 'package:stats':
## 
##     filter

## The following object is masked from 'package:graphics':
## 
##     layout

fig <- plot_ly(iot_tel,x = ~smoke,type = "histogram",color=~device, text='co', hovertext=~co)%>%layout(title='Distribution of smoke emmitted by devices')
fig

## Warning: `arrange_()` is deprecated as of dplyr 0.7.0.
## Please use `arrange()` instead.
## See vignette('programming') for more help
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_warnings()` to see where this warning was generated.

Plot_ly function

2. Visualize the co emission of the devices at different temperature for the days of the week.

days<-as.Date(as.POSIXct(iot_tel$ts, origin = '1970-01-01')) # get date from datetime
lwd<-weekdays(days) # extract day of the week
iot_tel$days<-lwd


library(ggplot2)

ggiot_tel<- ggplot(iot_tel, aes(temp,co,color=device))+geom_point()+facet_wrap(~days)+ 
  ggtitle('CO emissions of device at different temperature')+theme_dark()+
  xlab('Temperature')+ylab('CO emission')

ggiot_tel